Abstract: To solve problems of high risk, low efficiency and high defect omission rate in traditional manual inspection of wind turbine blades, and make up for shortcomings of existing UAV inspection technology in wind farm adaptability, stability and recognition accuracy, this paper develops a multi-rotor UAV-based intelligent inspection technology. An integrated framework of "carrier modeling-environment adaptation-path planning-defect detection" is constructed, integrating wind farm adaptive control, visual intelligent navigation and lightweight defect recognition technology. Accurate inspection is achieved by establishing UAV system model, hierarchical closed-loop control system, optimizing DWA navigation algorithm and designing YOLO-v5s lightweight model. Test results show that the technical scheme's performance data is significantly superior to traditional manual inspection, providing safe and efficient technical support for wind turbine blade operation and maintenance.